4.3. Screen and logfile output¶
As LAMMPS reads an input script, it prints information to both the screen and a log file about significant actions it takes to setup a simulation. When the simulation is ready to begin, LAMMPS performs various initializations, and prints info about the run it is about to perform, including the amount of memory (in MBytes per processor) that the simulation requires. It also prints details of the initial thermodynamic state of the system. During the run itself, thermodynamic information is printed periodically, every few timesteps. When the run concludes, LAMMPS prints the final thermodynamic state and a total run time for the simulation. It also appends statistics about the CPU time and storage requirements for the simulation. An example set of statistics is shown here:
Loop time of 0.942801 on 4 procs for 300 steps with 2004 atoms Performance: 54.985 ns/day, 0.436 hours/ns, 318.201 timesteps/s, 637.674 katom-step/s 195.2% CPU use with 2 MPI tasks x 2 OpenMP threads MPI task timing breakdown: Section | min time | avg time | max time |%varavg| %total --------------------------------------------------------------- Pair | 0.61419 | 0.62872 | 0.64325 | 1.8 | 66.69 Bond | 0.0028608 | 0.0028899 | 0.002919 | 0.1 | 0.31 Kspace | 0.12652 | 0.14048 | 0.15444 | 3.7 | 14.90 Neigh | 0.10242 | 0.10242 | 0.10242 | 0.0 | 10.86 Comm | 0.026753 | 0.027593 | 0.028434 | 0.5 | 2.93 Output | 0.00018341 | 0.00030942 | 0.00043542 | 0.0 | 0.03 Modify | 0.039117 | 0.039348 | 0.039579 | 0.1 | 4.17 Other | | 0.001041 | | | 0.11 Nlocal: 1002 ave 1006 max 998 min Histogram: 1 0 0 0 0 0 0 0 0 1 Nghost: 8670.5 ave 8691 max 8650 min Histogram: 1 0 0 0 0 0 0 0 0 1 Neighs: 354010 ave 357257 max 350763 min Histogram: 1 0 0 0 0 0 0 0 0 1 Total # of neighbors = 708020 Ave neighs/atom = 353.30339 Ave special neighs/atom = 2.3403194 Neighbor list builds = 26 Dangerous builds = 0
The first section provides a global loop timing summary. The loop time is the total wall-clock time for the simulation to run. The Performance line is provided for convenience to help predict how long it will take to run a desired physical simulation and to have numbers useful for performance comparison between different simulation settings or system sizes. The CPU use line provides the CPU utilization per MPI task; it should be close to 100% times the number of OpenMP threads (or 1 of not using OpenMP). Lower numbers correspond to delays due to file I/O or insufficient thread utilization.
The MPI task section gives the breakdown of the CPU run time (in seconds) into major categories:
Pair = non-bonded force computations
Bond = bonded interactions: bonds, angles, dihedrals, impropers
Kspace = long-range interactions: Ewald, PPPM, MSM
Neigh = neighbor list construction
Comm = inter-processor communication of atoms and their properties
Output = output of thermodynamic info and dump files
Modify = fixes and computes invoked by fixes
Other = all the remaining time
For each category, there is a breakdown of the least, average and most amount of wall time any processor spent on this category of computation. The “%varavg” is the percentage by which the max or min varies from the average. This is an indication of load imbalance. A percentage close to 0 is perfect load balance. A large percentage is imbalance. The final “%total” column is the percentage of the total loop time is spent in this category.
When using the timer full setting, an additional column is added that also prints the CPU utilization in percent. In addition, when using timer full and the package omp command are active, a similar timing summary of time spent in threaded regions to monitor thread utilization and load balance is provided. A new Thread timings section is also added, which lists the time spent in reducing the per-thread data elements to the storage for non-threaded computation. These thread timings are measured for the first MPI rank only and thus, because the breakdown for MPI tasks can change from MPI rank to MPI rank, this breakdown can be very different for individual ranks. Here is an example output for this section:
Thread timings breakdown (MPI rank 0): Total threaded time 0.6846 / 90.6% Section | min time | avg time | max time |%varavg| %total --------------------------------------------------------------- Pair | 0.5127 | 0.5147 | 0.5167 | 0.3 | 75.18 Bond | 0.0043139 | 0.0046779 | 0.0050418 | 0.5 | 0.68 Kspace | 0.070572 | 0.074541 | 0.07851 | 1.5 | 10.89 Neigh | 0.084778 | 0.086969 | 0.089161 | 0.7 | 12.70 Reduce | 0.0036485 | 0.003737 | 0.0038254 | 0.1 | 0.55
The third section above lists the number of owned atoms (Nlocal), ghost atoms (Nghost), and pairwise neighbors stored per processor. The max and min values give the spread of these values across processors with a 10-bin histogram showing the distribution. The total number of histogram counts is equal to the number of processors.
The last section gives aggregate statistics (across all processors) for pairwise neighbors and special neighbors that LAMMPS keeps track of (see the special_bonds command). The number of times neighbor lists were rebuilt is tallied, as is the number of potentially dangerous rebuilds. If atom movement triggered neighbor list rebuilding (see the neigh_modify command), then dangerous reneighborings are those that were triggered on the first timestep atom movement was checked for. If this count is non-zero you may wish to reduce the delay factor to ensure no force interactions are missed by atoms moving beyond the neighbor skin distance before a rebuild takes place.
If an energy minimization was performed via the minimize command, additional information is printed, e.g.
Minimization stats: Stopping criterion = linesearch alpha is zero Energy initial, next-to-last, final = -6372.3765206 -8328.46998942 -8328.46998942 Force two-norm initial, final = 1059.36 5.36874 Force max component initial, final = 58.6026 1.46872 Final line search alpha, max atom move = 2.7842e-10 4.0892e-10 Iterations, force evaluations = 701 1516
The first line prints the criterion that determined minimization was converged. The next line lists the initial and final energy, as well as the energy on the next-to-last iteration. The next 2 lines give a measure of the gradient of the energy (force on all atoms). The 2-norm is the “length” of this 3N-component force vector; the largest component (x, y, or z) of force (infinity-norm) is also given. Then information is provided about the line search and statistics on how many iterations and force-evaluations the minimizer required. Multiple force evaluations are typically done at each iteration to perform a 1d line minimization in the search direction. See the minimize page for more details.
If a kspace_style long-range Coulombics solver that performs FFTs was used during the run (PPPM, Ewald), then additional information is printed, e.g.
FFT time (% of Kspce) = 0.200313 (8.34477) FFT Gflps 3d 1d-only = 2.31074 9.19989
The first line is the time spent doing 3d FFTs (several per timestep) and the fraction it represents of the total KSpace time (listed above). Each 3d FFT requires computation (3 sets of 1d FFTs) and communication (transposes). The total flops performed is 5Nlog_2(N), where N is the number of points in the 3d grid. The FFTs are timed with and without the communication and a Gflop rate is computed. The 3d rate is with communication; the 1d rate is without (just the 1d FFTs). Thus you can estimate what fraction of your FFT time was spent in communication, roughly 75% in the example above.